The demographic and spatial distribution of inequality

I will consider issues around income distribution by examining three factors: the relationship between income, inequality and poverty; trends and levels of inequality; and decomposing inequality by income source.

The level of inequality

Depending on the data you use, our Gini coefficient is anywhere between 0.58 and 0.83.

There are many developing countries which are more equal than South Africa, and which have Gini coefficients of about 0.45 to 0.55. Developed countries tend to be even more equal, and have even lower Ginis. So South Africa's Gini coefficient is very high.

Because of differences in the data, and unreliable data derived from our surveys (see box on measuring inequality, page 17), it is difficult to reach certainty about trends.

However, most researchers believe intra-group inequality is rising, and inter-group inequality is falling. The latter has offset the former, so the net effect is a Gini coefficient that has, at best, been more or less stable since 1994, and may actually have risen.

Figure 1 shows the rise in intra-group inequality based on data drawn from the All Media and Products Survey (AMPS), a comprehensive survey administered by the South African Advertising Research Foundation. It shows that differences in intra-racial inequality remain significant and that all race groups are now more unequal than they were.

Figure 2 reflects the racial composition of 10 income deciles in 2005.

We can see that in 2005, 41 per cent of the top decile was black, confirming that, while race is still a major determinant of affluence, it no longer serves as the sole dividing line between the affluent and the rest of the population. The graph also confirms, however, that the poor remain overwhelmingly black. Other data show that the poor are also concentrated in rural areas.

The changing composition of the most affluent groups has resulted from new opportunities for some sections of the black population whose incomes were previously constrained by apartheid-era policies. This has led to rapid upward mobility in this population group, while the removal of the protection provided to some sections of the white population under apartheid has led to downward mobility for some.

Table 1 reflects the growth in the black share of the higher middle class and the working and lower middle classes since 1994, based on per capita earnings of R40 000 a year and R25 000 a year respectively (in 2000 terms). It shows that the black share of the higher middle class tripled over this period, and the black share of the working and lower middle classes more than doubled.

Decomposing inequality by income source

Individuals derive their incomes from various sources. Most comes from remuneration, but some from social transfers and other sources. Some government spending does not contribute directly to incomes, but has an effect on the aggregate distribution of resources by providing free or subsidised goods and services and therefore can play an important role in changing initial distributions.

Different sources of income may be more or less equitably shared. Incomes from private pensions, property rentals, and dividends paid to shareholders, for example, tend to be more unequally distributed than others, so their effect is to worsen overall levels of inequality.

By contrast, public spending, especially on transfers, is highly redistributive. At the height of apartheid, the state spent about nine times as much on each white person than on each black. This ratio has been reversed to the point where the state now spends about twice as much per capita on blacks than on whites. The main reason for that is that grant spending is strongly directed at the poorer parts of the population. Another reason is that most wealthier people opt out of some public services, such as health, which results in public spending in these sectors being almost entirely directed at the poor.

This shift in spending is one of the most dramatic ever seen outside a revolution, but it happened with very little opposition. But transfers and public spending comprise only a small proportion of all income, and wage inequality remains high. This is evident when comparing the Gini coefficient for all household income in South Africa to the Gini coefficient for wages only, as seen in Figure 3.

Wage inequality is quite high, and sets a floor for reducing inequality through the state.

In fact, even if we ignored inequalities in access to jobs, and assumed that all households had equal access to employment, social transfers, and other sources of income, wage inequality alone would still leave us with a Gini of 0.605, only a little lower than overall inequality.

What all this shows is that higher levels of employment do not necessarily reduce inequality. Obviously, greater employment plays a major role in reducing poverty, but because of high levels of wage inequality, it only has a minor effect on overall inequality.

This is because most inequality is driven by labour market outcomes, especially returns on education. This is a key reason why I would expect South African society to remain unequal for a long time to come. It also shows why one should avoid placing too much emphasis on inequality as such, and why one should focus on poverty instead.

Prof Servaas van der Berg is research chair: economics of social policy, University of Stellenbosch. This article is extracted from "Poverty and Inequality: Facts, trends, and hard choices", Edited proceedings of a Round Table convened by the Centre for Development and Enterprise, Number 15, August 2010. The full report can be accessed here.

Click here to sign up to receive our free daily headline email newsletter.